This article has been accepted for publication in a future issue of this journal, but has not been
fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2017.2674602, IEEE
Transactions on Industry Applications
Control of a Direct Matrix Converter with
Modulated Model Predictive Control
Manjusha Vijayagopal, Pericle Zanchetta, Lee Empringham, Liliana de Lillo, Luca Tarisciotti, Patrick Wheeler
The University of Nottingham, United Kingdom.
email:manjusha.vijayagopal@nottingham.ac.uk
Abstract—This paper investigates the use of a model- predic-
tive control strategy to control a direct matrix converter. The
proposed control method combines the features of the classical
Model Predictive Control and the Space Vector Modulation
technique into a Modulated Model Predictive Control. This new
solution maintains all the characteristics of Model Predictive
Control (such as fast transient response ,multi-objective control
using only one feedback loop, easy inclusion of non-linearities
and constraints of the system, the flexibility to include other
system requirements in the controller) adding the advantages of
working at fixed switching frequency and improving the quality
of the controlled waveforms. Simulation and experimental results
employing the control method to a direct matrix converter are
presented.
Index Terms—Matrix Converter, Model Predictive Control,
Modulated Model Predictive Control.
I. I NTRODUCTION
A direct matrix converter(MC) is a AC/AC converter that is
capable of converting varying amplitude, fixed frequency input
to varying amplitude and frequency output without employing
an intermediate DC-link capacitor. Nine bidirectional voltage-
blocking current-conducting switches arranged in the form of a
matrix as shown in Fig. 1 constitutes a MC making it possible
for bidirectional power flow. In the recent years, MCs have
reached a good level of technological maturity that allows
their practical and industrial implementation in a variety of
applications, such as industrial drives [1-2], power supplies
[3], aerospace applications [4-7]. MC is often known as an
all silicon converter as there are no bulky and heavy energy
storage devices [8].
The first matrix converter modulation strategy was proposed
by Alesina and Venturini [9]. Even though the initial strategy
proposed was capable of producing sinusoidal input and output
waveforms, the maximum voltage ratio it could achieve was
only 0.5. Later several other modulation techniques which
produce a higher voltage transfer ratio of 0.866 such as
Optimum Venturini Modulation and Space Vector Modulation
(SVM) were proposed later and SVM became the most widely
used modulation method for MC [10]. Since then, SVM
has been applied in many applications in conjunction with
feedback control strategies in order to regulate specific control
variables[11].
Model Predictive Control (MPC), introduced in the late
seventies [12] considers a model of the system in order
to predict its future behaviour over a time horizon. A cost
function represents the desired behaviour of the system. MPC
0093-9994 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
1
Bidirectional
switch
VA Matrix Converter
Vabc
IA SAa SAb SAc
VB
Input IB
Load
Filter IC SBa SBb SBc
VC
SCa SCb SCc
Fig. 1. Schematic of a direct matrix converter system
is an optimization problem where a sequence of future ac-
tuations is obtained by minimizing the cost function. It is
also referred to as a receding horizon control, which means
that at each instant the horizon is moved forward, the first
element of the sequence calculated at each step is applied at
that instant and all the calculation is repeated every sample
period [13]. Due to the high sampling rate used in the control
of power converters, solving the optimization problem of
MPC online is not practical. One approach is to use an
explicit solution of MPC, solving the optimization problem
offline. The resulting controller is a lookup table and can be
implemented without big computational effort. Considering
that power converters are systems with a finite number of
states, given by the possible combinations of the state of
the switching devices, the MPC optimization problem can be
simplified and reduced to the prediction of the behaviour of
the system for each possible state. Then, each prediction is
evaluated using the cost function and the state that minimizes
it, is selected. This approach has been successfully applied in
recent years for the control of power converters and drives,
like for current control and power control for three phase
Voltage Sourse Inverters [13-14] and for current and torque
control for induction motor and permanent magnet motor
drives[15-17]. Even though control of currents, torque and
flux are achieved, the switching frequency is variable and not
fixed. MPC presents several advantages, such as fast dynamic
response, easy inclusion of nonlinearities and constraints of
the system, flexibility to include other system requirements in
the controller, easy tuning of the control if the system model
is known and the possibility of modifying and extending the
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2017.2674602, IEEE
Transactions on Industry Applications
2

methodology depending on specific applications. M2PC. If M2PC needs to be implemented for MC feeding
One of the interesting features of MPC is that due to other loads such as an induction machine or a capacitive load,
the absence of a modulator, the control chooses and applies the load model needs to be derived appropriately. It is vital
one converter switching state for the entire sampling instant. to the performance of the controller that the load model is
This generates large ripples in the waveforms resulting in accurate.
variable and high switching frequencies compared to other The continuous time model of a resistive-inductive load is
control methods. Various methods have been proposed in the given by equation (3).
literature to improve the applied vector sequence [18] or to dio (t)
introduce a modulation scheme inside the MPC algorithm [19]. L = vo (t) − R io (t) (3)
dt
However these methods involve complicated expressions for
the switching time patterns and are not flexible enough to where R and L are the load resistance and inductance
include other system requirements in the cost function. This respectively,io (t) = [ia (t) ib (t) ic (t)]T the load currents and
is overcome by modulated model predictive control (M2PC) vo (t) = [va (t) vb (t) vc (t)]T the matrix converter output volt-
which has been proposed for a cascaded H-bridge converter age.
[20-22], active front-end rectifier [23], two level inverter [24], Discretising equation (3), using forward Euler approxima-
three phase rectifier [25], a neutral point clamp converter [26] tion, the discrete time model of the load can be obtained as
and indirect matrix converter [27]. Inspired from this new shown in equation (4).
approach, [28] elaborates the application of M2PC for a direct RTs Ts
matrix converter and includes experimental results to validate Io (k + 1) = (1 −
)Io (k) + Vo (k) (4)
L L
the simulation results discussed in [29]. This paper includes Equation (4) is then used to predict the load currents at
a comparison of the performance of M2PC with conventional the future sampling instants in order to formulate the cost
control methods such as MPC and PI controller. functions.

II. S YSTEM M ODEL

III. M ODULATED M ODEL P REDICTIVE C ONTROL FOR
In order to design the control system, an accurate model D IRECT M ATRIX C ONVERTER
is required. The model description can be divided into matrix
The M2PC strategy aims to combine the positive features
converter model and load model as presented in the following
of both SVM and MPC to obtain a model predictive control
subsections.
based algorithm with an intrinsic modulation scheme. A basic
control block diagram of the proposed strategy for a direct
A. Matrix converter model matrix converter is given in Fig. 2.
The voltages at the output of a matrix converter and input
Current reference,
currents are calculated from the input voltages and output Iref
currents respectively and can be derived directly from Fig. Io Io (k+2) G0-4
1. The voltages and currents are represented in terms of the Prediction of Cost function Duty cycle
load currents estimation calculation
switching functions related to each bidirectional switch in the
Vc t0-4
matrix converter as shown in equations (1) and (2). Constraints
Optimization of
switching
      sequence
va (t) SAa (t) SBa (t) SCa (t) vA (t)
 vb (t)  =  SAb (t) SBb (t) SCb (t)  . vB (t) Sij
(1) Converter voltage, Vc
Matrix
vc (t) SAc (t) SBc (t) SCc (t) vC (t) / abc
Converter

   T  
iA (t) SAa (t) SBa (t) SCa (t) ia (t) Load currents, Io
Load
iB (t) =  SAb (t) SBb (t) SCb (t)  .  ib (t)  (2) / abc
model
iC (t) SAc (t) SBc (t) SCc (t) ic (t)
where ia (t), ib (t) and ic (t) are the output currents, Fig. 2. Control block diagram for M2PC strategy
iA (t), iB (t) and iC (t) the input currents, va (t), vb (t) and vc (t)
the output voltages and vA (t), vB (t) and vC (t) the source A reference current is imposed upon the system and the
voltages and Sij (t) is the switching function for i= A,B and controller is designed using the system model so that the load
C and j= a,b and c. current tracks the reference. The measured currents are then
used to predict the value of current at (k+1). The reference
and predicted currents are then used to calculate the cost
B. Load model function which in turn is used to derive the duty cycles for
To determine the load current in the next sampling interval, the selected voltage vectors. Unlike 2-level inverters with two
a mathematical model of the load is required. Matrix con- active and three zero vectors, matrix converter usually makes
verters are usually connected to an inductive load and hence use of four active and three zero vectors to obtain sinusoidal
this paper considers the model of a RL load to demonstrate waveforms. Active vectors are those vectors that produce a

0093-9994 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2017.2674602, IEEE
Transactions on Industry Applications
3

non zero voltage. Zero vectors as the name suggests does not A. Prediction of Control Variables and Cost function calcula-
produce any voltage. tion
The M2PC utilizes the SVM vector sequence and calculates Similar to the MPC, the switching states are calculated
the duty cycles for each voltage vector based on the mini- based on a cost function minimization. The cost function can
mization of the cost function. Fixed switching frequency is include different performance factors according to the control
ensured in this case as the sequence of the vectors chosen by variables and to the constraints required. More than one vari-
the control will be applied within one sampling interval. The able can be controlled with the same control loop, providing
difference between a classical MPC and the M2PC is in the a multi-objective control approach and avoiding nested loops.
application time of the vectors. Unlike MPC where one vector In this paper, the M2PC algorithm is provisionally applied to
is applied for the whole sampling interval, in the M2PC at a direct matrix converter feeding a RL load to demonstrate
each sampling interval; four active and three zero vectors are current control capability.
selected by the cost function minimization algorithm and are For an RL load, the prediction of load current makes use
applied for their respective duty cycles. The active and zero of the current measurement at the present instant or instant
vectors are then arranged in a symmetrical manner as shown ’k’ and the load parameters as shown in (4). For MPC, the
in Fig. 3; to achieve minimum switching losses and reduce prediction of load current is carried out once using the optimal
harmonics. The switching pattern is similar to that used in switching state from the previous instant. Since the M2PC
a standard SVM scheme. t01 , t02 and t03 are the application incorporates a modulation scheme, more than one active vector
times for the three zero vectors and t1 , t2 , t3 and t4 for the will be applied during one sampling interval. For a direct
four active vectors. matrix converter, out of the 27 switching states, the six rotating
vectors are not considered for modulation resulting in 21
switching states to choose from. As in a standard SVM strategy
z3 V1 V2 z1 V3 V4 z2 V4 V3 z1 V 2 V1 z3 [10] it is assumed that a minimum of four active vectors is
required to get sinusoidal waveforms for a matrix converter.
t03 t1 t2 t01 t3 t4 t4 t3 t01 t2 t1 t03
2 2 2 2 2 2
t02
2 2 2 2 2 2 Hence the M2PC strategy will also consider four active vectors
Ts
and three zero vectors to apply during one sampling interval.
The predictions for four active vectors and zero vectors are
Fig. 3. Symmetrical switching pattern for a direct matrix converter done as if they were applied for the whole sampling interval.
For example, for one sequence containing four active and three
zero vectors, the predictions of load current are calculated as
Fig. 4 shows the flowchart of M2PC algorithm incorporating shown in equations (5) and (6). The predicted value of load
delay compensation for the computation. current for the three zero vectors will be the same and hence
it is only required to calculate once.
For active vectors :
RTs i Ts
Ioi (k + 1) = (1 − )Io (k) + Voi (k) (5)
L L
for i=1,2,..4.
For zero vectors:
RTs
Io (k + 1) = (1 −)Io (k) (6)
L
Once the currents at (k+1) instant are predicted, (k+2)
predictions can be made by substituting Io (k) with Io (k +1)in
equations (5) and (6).
Just as in MPC, the switching states are selected based on
a cost function minimization. The cost function can include
different performance factors according to the control variables
and constraints required. More than one variable can be
controlled with the same control loop, providing a multi-
objective control approach and avoiding nested loops.
The cost function for load current control is essentially the
error between the current demand and predicted current at
(k+2) instant if the computation delay is compensated. This
results in five cost functions G0 , G1 , G2 , G3 and G4 for each
active and zero vectors. The set of quadratic cost functions
can be calculated using the pre-calculated predicted currents
Fig. 4. Flowchart showing the M2PC algorithm in equation (5) as shown in (7).
Gi = (Ioref (k + 2) − Ioi (k + 2))2 (7)

0093-9994 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2017.2674602, IEEE
Transactions on Industry Applications
for i=0,1,...4.
where Ioref (k + 2) is the current demand at (k+1) instant.
The final cost function used for the minimization algorithm
constitutes the cost functions for the active and zero vectors
and is expressed as shown in equation (8). The minimization
algorithm is run for 18 combinations of active and zero vector
sequences.
G(f inal) = G1 d1 + G2 d2 + G3 d3 + G4 d4 + G0 d0
where d0 , d1 , d2 , d3 and d4 are the duty cycles for the voltage
vectors.
Once the cost functions for each of the active and zero vec-
tors are calculated; the predictive control will choose the best
sequence of vectors which produces the lowest error. These
vectors are then applied within the sampling interval with their
respective duty cycles. The duty cycles are calculated based
on the cost functions values of the active and zero vectors.
The multi-objective control capability of the M2PC can be
achieved by simply adding the control variables to the cost
function for each vector and by specifying a weighting factor
for each of them. In this manner, it is possible to also consider
certain constraints for the optimal operation of the system.
This is demonstrated in [30] by controlling the stator currents
and input reactive power of an induction motor fed by a direct
matrix converter without the use of nested loops.
B. Calculation of Duty Cycles
The application times for each of the vectors are calculated
from the cost functions computed for each the switching state.
This method is based on the assumption that per sampling
interval, the system behavior is linear in nature. From this it
is possible to derive the application times for each vector as
a percentage of the total sampling time. The relation between
the application times of the vectors and their corresponding
cost functions can be written as follows.
k1
t= (9)
G The M2PC is applied to a direct matrix converter feeding
where k1 is the proportional constant and G is the cost
function. Using the above equation for four active vectors and
zero vector for a MC will result in a set of five equations, each
dictating the application times for a particular vector in terms
of the proportional constant. The condition that holds true for
all the vectors applied is that the sum of the application times
for all the vectors should be equal to the sampling time as
shown in (10).
t1 + t2 + t3 + t4 + t0 = Ts
where t1 , t2 , t3 , t4 are the times for the four active vectors and
t0 is the time for three zero vectors. By substituting equation
(9) in (10), the proportional constant can be obtained as shown
in (11).
G0 G1 G2 G3 G4 Ts
k1 =
A
where A = (G1 G2 G3 G4 + G0 G2 G3 G4 + G0 G1 G3 G4 +
G0 G1 G2 G4 + G0 G1 G2 G3 ).
0093-9994 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
4
The application times for the vectors can then be derived
by substituting the value of the proportional constant. The
resulting set of duty cycles is shown in equation (12).
(G0 G2 G3 G4 Ts )
t1 =
A
(G0 G1 G3 G4 Ts )
t2 =
A
(8) (12)
(G0 G1 G2 G4 Ts )
t3 =
A
(G0 G1 G2 G3 Ts )
t4 =
A
Once the application times for active vectors are obtained,
the time for zero vectors can be calculated from equation (10).
t0 = Ts − (t1 + t2 + t3 + t4 ) (13)
The duty cycles for each of the vector will be related to the
error associated with that voltage vector. The direct relation
of the duty cycles with the cost functions makes this method
unique. The active and zero vectors are then applied for their
respective duty cycles within one sampling interval.
The active and zero vectors are then arranged in a symmetri-
cal manner as shown in Fig. 3; to achieve minimum switching
losses and reduced harmonics. The switching pattern is similar
to that used in a SVM scheme. t01 , t02 and t03 are the appli-
cation times for the three zero vectors and t1 , t2 , t3 and t4 for
the four active vectors in Fig. 3.
The switching frequency that can be attained using this
method depends on the computational speed of the micro-
processor being used. The matrix converter requires a sig-
nificant computational overhead for this method since 27
switching states are available. The maximum controller update
frequency obtained for M2PC within the laboratory when
using a Texas Instruments C6713 DSP was 20kHz.
IV. S IMULATION R ESULTS
an RL load and simulations are done in Matlab Simulink
environment to study its performance. A sampling time of
80µs is considered. The simulation also takes into account the
four-step commutation in matrix converter and the measure-
ment delays in the control platform. An LC input filter with a
damping resistor is employed to mitigate the high frequency
components of the input current of the matrix converter. A
control block diagram showing the different steps involved
in the load current control of a direct matrix converter using
(10) M2PC is shown in Fig. 5.
(11)
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2017.2674602, IEEE
Transactions on Industry Applications
5

Matrix
Converter of switching frequency, 12.5kHz and its multiples. This is
VA IA
Va a result of fixed switching frequency operation. To analyse
VB IB Input Vb
Load
the quality of the controlled waveforms the Total Harmonic
filter
VC IC Vc
Distortion (THD) of Phase A load current is calculated and
Io is approximately 6.3%. To demonstrate the fast dynamic
Sij abc to
response of M2PC, a step change from 2A to 4A and then
Cost function to 2A. The resulting waveform is shown in Fig. 7 (top).
Iref minimization
Prediction of
load currents
Duty cycle calculation
Io(k+1)

load currents (A)

5

Alpha−beta
Fig. 5. Control block diagram of M2PC for a matrix converter
0
The system parameters for both the simulation and experi- −5 Ialpha Ibeta Irefalpha Irefbeta
mental tests are given in Table. I.
0 0.05 0.1 0.15 0.2 0.25 0.3
TABLE I
S YSTEM PARAMETERS

load currents (A)

Parameter Value Unit 5

Alpha−beta
0
Filter inductance 0.7 mH
Filter capacitance (delta) 8.3 µF −5 Ialpha Ibeta Irefalpha Irefbeta
Damping resistor 15 Ω
Load resistance 10 Ω 0 0.05 0.1 0.15 0.2 0.25 0.3
Time (s)
Load inductance 3.75 mH
Sampling time 80 µs
Fig. 7. Alpha-beta currents when the frequency of reference is changed

The load current of a matrix converter feeding an RL load is It is evident from the figure that the load current im-
controlled using M2PC and the resulting waveforms are shown mediately follows the step demand in reference current and
below. A load current reference of 5A at 30Hz is demanded reaches steady state immediately. This indicates that M2PC
from the system. Fig. 6 shows the controlled three phase load can provide fast transient response which is a characteristic it
currents, MC line voltage at steady state and the Fast Fourier inherited from MPC. In addition to the step change in current
Transform (FFT) of the load current at steady state. reference magnitude, a step change in the frequency of the
current reference from 20Hz to 40Hz is also introduced.The
resulting waveform is shown in Fig. 7 (bottom). The results
load currents(A)

Phase a indicate that M2PC is able to handle any abrupt changes in

5
Three ph

Phase b the reference signal without any overshoots or instability.

0
Phase c In order to conduct a quantitative analysis of the transient
−5
response of the M2PC controller, load current control of matrix
0 0.02 0.04 0.06 0.08 0.1 converter is implemented using MPC and a conventional
line voltage (V)

200 PI Controller. The system parameters for the three methods

MC Output

remain constant. A step demand in load current amplitude

0 from 2A to 4A is applied to the controller and the resulting
d-axis load currents with the reference signal for all the three
−200 control methods is shown in Fig. 8. To conduct a fair analysis
0 0.02 0.04 0.06 0.08 0.1
Time (s) of the transient response of the controllers considered, the
Magnitude(A)

PI controller is tuned to achieve very fast response to set a

0.2 benchmark for the comparison.
0.1 The simulation results for load current control of a direct
0 MC with RL load are considered to compare the performance
0 0.5 1 1.5 2 2.5 3 of M2PC controller with PI Controller and MPC. The THD of
Frequency(Hz) 4
x 10 the currents controlled by the three methods can be considered
as a measure of their steady state performance. To analyse the
Fig. 6. Three phase load currents of MC controlled by M2PC transient performance of the control strategies, the rise time
of the controlled currents during a step change is considered.
It is worth noting the presence of harmonics in the range Rise time is the time taken by the control variable to reach

0093-9994 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2017.2674602, IEEE
Transactions on Industry Applications
6

from 10% to 90% of the steady state value. A comparison of

the THDs and rise times for load current control of a MC for
different controllers such as PI controller, MPC and M2PC
from simulation results is tabulated in Table. II.

8 Idref Id
6 M2PC
4
2
0
d−axis load current (A)

0.04 0.045 0.05 0.055 0.06 0.065

8
6 MPC
4
2
0 Fig. 9. Experimental setup
0.04 0.045 0.05 0.055 0.06 0.065
8
6 PI Controller The experimental setup shown in Fig. 9 includes a matrix
4 converter using SK60GM123 IGBT modules rated at 1200V
2 and 60A, a current direction detection circuit for four step
0 commutation and a clamp circuit for over voltage protection.
0.045 0.04
0.05 0.055 0.06 0.065 Each IGBT module consists of two diodes and two anti-
Time (s)
parallel IGBT connected in the common emitter configuration.
There are three current sensors to measure the output currents
Fig. 8. d-axis load current with reference during a step change in the
amplitude and three voltage sensors to measure the supply voltage to
the converter. The control platform includes a Texas instru-
ment DSP and a FPGA card developed at Power Electronics
TABLE II Machines and Control (PEMC) Group, The University Of
C OMPARISON OF THD S AND RISE TIMES OF LOAD CURRENTS FOR Nottingham. The M2PC algorithm is implemented on the DSP
DIFFERENT CONTROL STRATEGIES
at a sampling frequency of 80µ s. An FPGA interface with the
DSP ensures the four-step commutation of the switches. The
Steady State T ransient P erf ormance
Control Strategy setup is powered using Chroma Programmable AC source as
P erf ormance(T HD) (Risetime)
indicated in Fig. 9.
The load currents of the direct matrix converter are con-
P I Controller 1.89% 4.1ms trolled using the M2PC strategy. A reference current of 5A
MP C 8.09% 0.34ms at 30Hz is demanded from the system and the experimental
M 2P C 6.3% 0.65ms results of controlled load currents, MC output line voltage and
the harmonic spectrum of Phase A load current are shown
in Fig. 10. It is evident from Fig. 10 that the load currents
From Table. II it can be seen that PI Controller with SVM
are sinusoidal and reaches the steady state value without any
resulted in waveforms with least harmonic distortion followed
error. The THD of the controlled waveforms is approximately
by M2PC. MPC resulted in waveforms with the worst current
10.8%. The harmonic spectrum of Phase A load current reveals
quality. Predictive control based algorithms such as MPC and
harmonics in the range of switching frequency (12.5kHz)
M2PC had faster dynamic response with rise times of 0.34ms
and its multiples which confirms fixed switching frequency
and 0.65ms respectively; compared to that of PI controller
operation. This validates the simulation results shown in Fig.
that resulted in a rise time of 4.1ms. These results prove that
6.
M2PC is capable of fast dynamic response very similar to
To analyse the transient behavior of the control strategy, a
that of MPC when compared to a traditional PI controller and
step demand in the amplitude and frequency of the reference
deliver waveforms with enhanced current quality compared to
current waveform is applied. A step in the amplitude of
MPC.
reference current waveform from 2A to 4A is applied and
the frequency of the reference load current is changed from
V. E XPERIMENTAL R ESULTS
20Hz to 40Hz and the resulting waveforms are shown in Fig.
To validate the simulation results discussed in the previous 11.
section, the M2PC is implemented on a matrix converter
feeding an RL load in the laboratory. The parameters of the
system are given in Table. I.

0093-9994 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2017.2674602, IEEE
Transactions on Industry Applications
7

load currents(A) the effect of sampling time on the quality of the controlled
waveforms in this case the load currents, the M2PC strategy
Three phase

5 Phase a
Phase b is implemented for three different sampling times such as
0 50µs, 80µs and 100µs. The quality of the controlled currents
Phase c
−5 are assessed by comparing the THDs for these sampling times.
0 0.02 0.04 0.06 0.08 0.1 Both simulation and experimental tests are conducted for three
different sampling intervals or switching frequencies and the
line voltage (V)

200
MC Output

THDs of the resulting current waveforms are tabulated as

0 shown in Table. III.

−200 TABLE III

0 0.02 0.04 0.06 0.08 0.1 C OMPARISON OF THD S FOR M2PC WITH DIFFERENT SAMPLING
Time (s) INTERVALS
Magnitude(A)

Total Harmonic Distortion

Sampling time
0.1 Simulation Results Experimental Results

0 50µs 4.0% 7.33%

0 0.5 1 1.5 2 2.5 3
Frequency(Hz) 4 80µs 6.3% 10.8%
x 10
100µs 7.5% 12.07%

Fig. 10. Phase A load current, MC output line voltage and harmonic spectrum
of load current at steady state Table. III indicates that there is a considerable effect on the
quality of the controlled waveforms as the sampling interval
is changed. The current quality was the best when a sampling
time of 50µs is considered and the worst when the sampling
load currents(A)

5 Phase a time is 100µs. It is interesting to note that the M2PC strategy

Three phase

Phase b is also dependant on the sampling interval considered which

Phase c
0 is expected as it is predominantly a predictive control based
method.
-5
0 0.05 0.1 0.15 0.2 VII. C ONCLUSION
A control method with the features of MPC and a mod-
load currents (A)

5
ulation scheme similar to SVM is proposed in this paper
for a direct matrix converter drives. The ability to control
Alpha-beta

0 different parameters simultaneously, high controller bandwidth

-5 Irefalpha Ialpha Irefbeta Ibeta
0.1 0.2 0.3
Time (s)
Fig. 11. Dynamic performance of M2PC during step change in the magni-
tude(top) and frequency(bottom) of the reference load current
From Fig. 11, it is evident that the load currents respond
to the sudden changes in amplitude and frequency instanta-
neously without any delay. The results from the experimental
tests conforms to the simulation results which proves the
system model and control strategy. To summarize, the load
current control achieved by this method is characterized by
very fast transient response as in the case of MPC and an
improved steady state response due to the presence of an
inbuilt modulation scheme.
VI. E FFECT OF S AMPLING T IME ON WAVEFORM Q UALITY
Since the performance of MPC is largely dependant on
the sampling interval, it is worth studying if this applies to
the M2PC. For MPC, quality of the controlled currents are
improved as the sampling interval gets smaller. To analyse
and constant switching frequency are the main highlights of
this method. The constant switching behavior of the SVM
0.4 0.5
utilised in this method guarantees that a predictive based
control method can now be used where the traditional associ-
ated problems with input filter sizing, harmonic performance,
switching loss and hence thermal management design can now
be addressed in a more predictable and systematic way. A
comparison of the performance of M2PC with control methods
such as MPC and PI controller showed that M2PC is capable
of fast dynamic response. The experimental results presented
in this paper validates the M2PC strategy for load current
control of a direct matrix converter.
VIII. R EFERENCES
[1] Wheeler, P. W., J. C. Clare, M. Apap, D. Lampard, S.
J. Pickering, K. J. Bradley, and L. Empringham. ”An
integrated 30kw matrix converter based induction motor
drive.” In 2005 IEEE 36th Power Electronics Specialists
Conference, pp. 2390-2395. IEEE, 2005.
[2] P. W. Wheeler, J. C. Clare, D. Katsis, L. Empringham,
M. Bland and T. Podlesak, ”Design and construction
of a 150 KVA matrix converter induction motor drive,”
Power Electronics, Machines and Drives, 2004. (PEMD

0093-9994 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2017.2674602, IEEE
Transactions on Industry Applications
2004). Second International Conference on (Conf. Publ.
No. 498), Edinburgh, UK, 2004, pp. 719-723 Vol.2.
[3] P. Zanchetta, P. W. Wheeler, J. C. Clare, M. Bland, L.
Empringham and D. Katsis, ”Control Design of a Three-
Phase Matrix-Converter-Based ACAC Mobile Utility
Power Supply,” in IEEE Transactions on Industrial Elec-
tronics, vol. 55, no. 1, pp. 209-217, Jan. 2008.
[4] L. Empringham, J. W. Kolar, J. Rodriguez, P. W.
Wheeler, and J. C. Clare, ”Technological Issues and
Industrial Application of Matrix Converters: A Review,”
IEEE Transactions on Industrial Electronics, vol. 60, pp.
4260-4271, 2013.
[5] L. de Lillo, L. Empringham, P. Wheeler, J. Clare and K.
Bradley, ”A 20 KW matrix converter drive system for
an electro-mechanical aircraft (EMA) actuator,” Power
Electronics and Applications, 2005 European Confer-
ence on, Dresden, 2005, pp. 6 pp.-P.6.
[6] Wheeler, P. W., J. C. Clare, M. Apap, L. Empringham,
L. De Lilo, K. Bradley, C. Whitley, and G. Towers.
”An electro-hydrostatic aircraft actuator using a matrix
converter permanent magnet motor drive.” In Power
Electronics, Machines and Drives, 2004.(PEMD). Sec-
ond International Conference on (Conf. Publ. No. 498),
vol. 2, pp. 464-468. IET.
[7] Lee Empringham, L. de Lillo, P. W. Wheeler and J. C.
Clare, ”Matrix Converter Protection for More Electric
Aircraft Applications,” IECON 2006 - 32nd Annual
Conference on IEEE Industrial Electronics, Paris, 2006,
pp. 2564-2568.
[8] J. Rodriguez, M. Rivera, J. W. Kolar, and P. W. Wheeler,
”A Review of Control and Modulation Methods for
Matrix Converters,” IEEE Transactions on Industrial
Electronic, vol. 59, pp. 58-70, 2012.
[9] A. Alesina and M. Venturini, ”Analysis and design of
optimum-amplitude nine-switch direct AC-AC convert-
ers,” IEEE Transactions on Power Electronics, vol. 4,
pp. 101-112, 1989.
[10] D. Casadei, G. Grandi, G. Serra, and A. Tani, ”Space
vector control of matrix converters with unity input
power factor and sinusoidal input/output waveforms,” in
Fifth European Conference on Power Electronics and
Applications, 1993, pp. 170-175 vol.7.
[11] P. Wheeler, J. C. Clare, M. Apap, D. Lampard, S.
Pickering, K. J. Bradley, et al., ”An Integrated 30kW
Matrix Converter based Induction Motor Drive,” in IEEE
36th Power Electronics Specialists Conference, 2005.
PESC ’05, pp. 2390-2395.
[12] E. F. Camacho and C. B. Alba, Model predictive control:
Springer, 2013.
[13] J. Rodriguez and P. Cortes, Predictive control of power
converters and electrical drives vol. 37: John Wiley and
Sons, 2012.
[14] J. Rodriguez, J. Pontt, C. A. Silva, P. Correa, P. Lezana,
P. Cortes, et al., ”Predictive Current Control of a Volt-
age Source Inverter,” IEEE Transactions on Industrial
Electronics, vol. 54, pp. 495-503, 2007.
[15] J. Rodriguez, J. Pontt, R. Vargas, P. Lezana, U. Ammann,
P. Wheeler, et al., ”Predictive direct torque control of
0093-9994 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
8
an induction motor fed by a matrix converter,” in Euro-
pean Conference on Power Electronics and Applications,
2007, pp. 1-10.
[16] P. Correa, M. Pacas, and J. Rodriguez, ”Predictive
Torque Control for Inverter-Fed Induction Machines,”
IEEE Transactions on Industrial Electronics, vol. 54, pp.
1073-1079, 2007.
[17] T. Geyer, G. A. Beccuti, G. Papafotiou, and M. Morari,
”Model Predictive Direct Torque Control of permanent
magnet synchronous motors,” in IEEE Energy Conver-
sion Congress and Exposition (ECCE), 2010, pp. 199-
206.
[18] H. Jiabing and Z. Q. Zhu, ”Improved Voltage-Vector
Sequences on Dead-Beat Predictive Direct Power Con-
trol of Reversible Three-Phase Grid-Connected Voltage-
Source Converters,” IEEE Transactions on Power Elec-
tronics, vol. 28, pp. 254-267, 2013.
[19] P. Antoniewicz and M. P. Kazmierkowski, ”Virtual-
Flux-Based Predictive Direct Power Control of AC/DC
Converters With Online Inductance Estimation,” IEEE
Transactions on Industrial Electronics, , vol. 55, pp.
4381-4390, 2008.
[20] L. Tarisciotti, P. Zanchetta, A. Watson, J. Clare, and
S. Bifaretti, ”Modulated Model Predictive Control for
a 7-Level Cascaded H-Bridge back-to-back Converter,”
IEEE Transactions on Industrial Electronics, vol. PP, pp.
1-1, 2014.
[21] L. Tarisciotti, P. Zanchetta, A. Watson, J. Clare, S. Bi-
faretti and M. Rivera, ”A new predictive control method
for cascaded multilevel converters with intrinsic mod-
ulation scheme,” Industrial Electronics Society, IECON
2013 - 39th Annual Conference of the IEEE, Vienna,
2013, pp. 5764-5769.
[22] L. Tarisciotti, P. Zanchetta, A. Watson, P. Wheeler, J. C.
Clare and S. Bifaretti, ”Multiobjective Modulated Model
Predictive Control for a Multilevel Solid-State Trans-
former,” in IEEE Transactions on Industry Applications,
vol. 51, no. 5, pp. 4051-4060, Sept.-Oct. 2015.
[23] L. Tarisciotti, P. Zanchetta, A. Watson, J. Clare, M.
Degano, and S. Bifaretti, ”Modulated model predictve
control (M2PC) for a 3-phase active front-end,” in IEEE
Energy Conversion Congress and Exposition (ECCE), ,
2013, pp. 1062-1069.
[24] Rivera, M., F. Morales, C. Baier, J. Munoz, L. Tarisciotti,
P. Zanchetta, and P. Wheeler. ”A modulated model pre-
dictive control scheme for a two-level voltage source in-
verter.” In Industrial Technology (ICIT), 2015 IEEE In-
ternational Conference on, pp. 2224-2229. IEEE, 2015.
[25] L. Tarisciotti, P. Zanchetta, A. Watson, J. C. Clare, M.
Degano and S. Bifaretti, ”Modulated Model Predictive
Control for a Three-Phase Active Rectifier,” in IEEE
Transactions on Industry Applications, vol. 51, no. 2,
pp. 1610-1620, March-April 2015.
[26] Rivera, M., M. Perez, V. Yaramasu, B. Wu, L. Tarisciotti,
P. Zanchetta, and P. Wheeler. ”Modulated model predic-
tive control (M 2 PC) with fixed switching frequency
for an NPC converter.” In 2015 IEEE 5th International
Conference on Power Engineering, Energy and Electri-
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TIA.2017.2674602, IEEE
Transactions on Industry Applications
9

cal Drives (POWERENG), pp. 623-628. IEEE, 2015.

[27] M. Rivera, C. Uribe, L. Tarisciotti, P. Wheeler and
P. Zanchetta, ”Predictive control of an indirect matrix
converter operating at fixed switching frequency and
unbalanced AC-supply,” 2015 IEEE International Sym-
posium on Predictive Control of Electrical Drives and
Power Electronics (PRECEDE), Valparaiso, 2015, pp.
38-43.
[28] M. Vijayagopal, P. Zanchetta, L. Empringham, L. De
Lillo, L. Tarisciotti and P. Wheeler, ”Modulated model
predictive current control for direct matrix converter
with fixed switching frequency,” Power Electronics and
Applications (EPE’15 ECCE-Europe), 2015 17th Euro-
pean Conference on, Geneva, 2015, pp. 1-10.
[29] M. Vijayagopal, L. Empringham, L. de Lillo, L.
Tarisciotti, P. Zanchetta and P. Wheeler, ”Control of
a direct matrix converter induction motor drive with
modulated model predictive control,” 2015 IEEE Energy
Conversion Congress and Exposition (ECCE), Montreal,
QC, 2015, pp. 4315-4321.
[30] M. Vijayagopal, L. Empringham, L. de Lillo, L.
Tarisciotti, P. Zanchetta and P. Wheeler, ”Current control
and reactive power minimization of a direct matrix
converter induction motor drive with Modulated Model
Predictive Control,” 2015 IEEE International Sympo-
sium on Predictive Control of Electrical Drives and
Power Electronics (PRECEDE), Valparaiso, 2015, pp.
103-108.

0093-9994 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.